pugscode.org/ planetsix.perl.org/ | nopaste: sial.org/pbot/perl6 | evalbot: perl6: say 3; (or rakudo:, pugs:, elf:, etc) | irclog: irc.pugscode.org/ | ~330 days 'til Xmas
Set by mncharity on 25 January 2009.
00:08 justatheory left 00:20 iblechbot left 00:22 Chillance joined
shinobi_cl i know that might be too early.. 00:33
but are there plans for a parrot widget library yet? 00:34
diakopter is a widget something a parrot eats? 00:35
shinobi_cl i've seen some SDL code for parrot somewhere. If parrot can digest that, probably can digest a widget too. i guess 00:36
but, im not sure if it has to be done... If i make, for example, a simple library that allows to show windows, labels and buttons on perl6.... 00:45
That would allow any language that parrot supports to make use of it? 00:46
pugs_svn r25239 | hinrik++ | [util/perl6.vim] handle =END/=for END/=begin END correctly 00:47
01:00 justatheory joined 01:01 M_o_C left 01:07 c9s left 01:10 ab5tract joined 01:16 bacek_ left 01:18 DemoFreak left 01:32 Guest12310 left 01:52 braceta_ joined 02:03 shinobi_cl left 02:04 ab5tract left 02:07 braceta left 02:14 braceta_ left 02:59 justatheory left 03:15 shinobi_cl joined 03:21 Limbic_Region left 03:22 Tene_ joined 03:36 Tene left 03:37 shinobi_cl left 03:59 justatheory joined 04:00 justatheory left 04:09 wknight8111 joined 04:29 c9s joined 04:42 gravity joined
gravity rakudo's Makefile's spectest target has a minor bug. The URL for pugscode is missing the second slash after http 04:43
04:45 SamB left 04:56 dalek left, dalek joined 04:58 wknight8111 left 05:05 Chillance left 05:12 finanalyst joined 05:38 gravity left 05:40 c9s_ joined 05:57 c9s left, c9s_ is now known as c9s 06:14 alester joined, alester left 06:57 c9s_ joined 07:04 PantheraPardus joined 07:09 c9s_ left 07:10 c9s_ joined 07:16 c9s left, c9s_ is now known as c9s 07:56 c9s_ joined 07:58 eternaleye left 08:01 c9s__ joined 08:03 eternaleye joined 08:07 c9s left, c9s__ is now known as c9s 08:18 c9s_ left 08:41 finanalyst left 09:00 maerzhase joined 09:44 barney joined 09:54 iblechbot joined 09:55 M_o_C joined
samlh I was just bit by the spectest target bug myself 10:07
10:13 pmurias joined, ejs joined 10:23 oskie left 10:29 DemoFreak joined 11:10 ejs left 11:17 ayrnieu_ joined 11:27 ayrnieu left 11:32 Muixirt joined 11:45 kane__ left 11:50 ejs joined 11:51 ejs left 11:56 smtms left 12:05 pmurias left 12:14 ejs joined 12:21 PantheraPardus left 12:41 mberends joined 12:48 maerzhase1 joined, maerzhase left 13:10 ejs left 13:15 smtms joined 13:28 Whiteknight joined 13:46 shinobi_cl joined 14:01 SamB joined 14:09 meppl joined 14:12 masak joined 14:24 wknight8111 joined 14:25 Whiteknight left, shinobi_cl left
masak mberends: hiya. 14:26
mberends HAI
masak :)
right now, my goal is simply a no-design-up-front packager that handles dependencies. 14:27
it's being built at exactly the right moment, because I'm planning to add my first dependency shortly. 14:28
and after that, more will arise.
for example Druid -> SVG.
szabgab is doing something which depends on HTML::Template. so far he's content with modding @INC at BEGIN time, so he doesn't feel the need for proto. yet. :) 14:29
szabgab I can hear you !
masak szabgab: :)
I don't mind. 14:30
14:30 ascent_ joined
masak szabgab: thanks for the nice emails on the november-wiki list, btw. 14:30
and szabgab++ for the CPAN module.
szabgab have you seen my mail ? it seems that the release to CPAN and install via CPAN.pm works
masak haven't looked at it yet, but its mere existence is really good news.
szabgab: will try it shortly.
szabgab yeah and Andreas fixed PAUSE so it was now indexed 14:31
masak that's good.
szabgab there are a few thing we need to formalize to make it easier for everyone else to upload modules 14:32
masak indeed.
I can enumerate the things I will expect of a module in proto:
szabgab one thing Ihave not touched yet is how CPAN.pm will be able to check if the prereqs of a package are met?
brb
masak 1. a Makefile.PL
2. a lib/ directory (so November is currently disqualified) 14:33
...and that's it.
mberends is that @INC at BEGIN time a short term hack or an official technique? Perl 5 had a reason to disallow it and force 'use lib' instead. 14:34
masak if a project wants to declare dependencies, it has to have a file 'deps.proto' in the same dir that contains the above files.
mberends: I don't mind it, but it only works if you don't publish stuff.
szabgab which BEGIN block do you mean ?, in which file?
masak (because the path to the thing depended on is hard-coded)
szabgab: I don't know, it's your trick. :)
szabgab oh you don't need that any more assuming you have PERL6LIB configured correctly 14:35
masak that's right.
but even setting PERL6LIB manually is a chore when you have many projects. 14:36
mberends the drawback of PERL6LIB is that your program's parent process needs to set it
masak proto will set it for you.
mberends: I know. it's a pity.
I've been thinking a lot about that in the last few days.
here's my current solution:
(not yet implemented)
szabgab I have PERL6LIB set in my .bashrc to the @INC of my perl5 14:37
actually not even that, as I am installing all my modules in a private directory
masak stop the script, print out a message to STDERR saying "hey, I need you to set these environment variables for me, here's how you do it". next time the script starts, check if they've been set. otherwise, explain again what needs to be done.
I plan to use this procedure for PARROT_DIR as well as PERL6LIB. it feels like a reasonable compromise between safety and convenience. 14:38
szabgab and right now I still assume PARROT_DIR point to trunk and rakudo being in languages/rakudo
masak me too. will have to look at that. 14:39
proto needs to be able to handle all such setups.
also, it needs to be able to respect that some things are checked out from github with the commit bit set.
mberends moritz and I have switched to Parrot inside Rakudo. Second option in perlgeek.de/blog-en/perl-6/where-ra...lives.html 14:40
szabgab yeah it will be nice if we can easily mix installed packages with development packages from git
masak mberends: makes sense for Rakudo devs, I guess.
szabgab: how do you mean? 14:41
mberends also make sense for Rakudo users who are not interested in Parrot
masak aye.
szabgab masak, I dont know :-)
masak mberends: makes me wonder if I shouldn't make that installation method the default in proto... 14:42
szabgab mberends, so the rakudo in languages/rakudo limitations was lifted
mberends yes
pmichaud++ made that second case possible, I do not think the two trees can be at any arbitrary locations. 14:43
masak that's my understanding too.
and moritz_ reported problem with 'make spectest' with his setup. 14:44
szabgab anyway I think the best might be to require a PARROT_DIR and a RAKUDO_DIR variable now
masak szabgab: aye.
szabgab yeah, so I'll update mu code accordingly 14:45
masak with the latter defaulting to $PARROT_DIR/languages/rakudo or perhaps $PARROT_DIR/..
szabgab I am not sure I want that default 14:46
masak hm, maybe not.
mberends consider how we 'git clone' each other's projects all over the filesystem, every project adds another directory to PERL6LIB. That's unaccaptable. The usual Unix remedy is symbolic directory links, and tell the Windows users to try Cygwin.
szabgab though we can be nice and guess (or actually check if either of thoes exist)
masak mberends: why unacceptable?
szabgab but on the other hand it is not that difficult for the user to configure both
masak mberends: that's exactly the problem that I want proto to solve. 14:47
the managing of PERL6LIB.
szabgab: right. perhaps just require that, then.
szabgab mberends, I don't think telling the windows users anything is will work
mberends no, it hasn't up to now.
unacceptable in that scaling to hundreds of modules (just watch me) will make the design cumbersome. 14:48
how about 'git clone' into a temp area and then proto moves the results to a common lib location. CPAN does that. 14:49
masak mberends: works for me. it's just not a high priority right now. 14:50
though I'm getting a bit bothered myself by proto's soiling its own directory. thinking of creating a directory 'projects' or whatever, and .gitignoring it. 14:51
mberends I like it enough to work on it, but not today. SVG::Tiny has a cold, I must cure that first.
masak mberends++ 14:52
I might get back to SVG today. I have some plans.
mberends and Pod::Parser risks bitrot through lack of love
masak but November also has a cold.
mberends achoo! 14:53
masak and before that, I need to write a small essay on Chinese Literature.
mberends ok, off to work everyone, chat again later!
masak aye aye, sir! 14:54
14:58 rhr joined 15:01 Nom- joined
Nom- Hi all :) 15:03
15:05 icwiener joined
Nom- Can anyone suggest a good guide for getting started? I haven't kept up with perl6, but want to catch up now that we're weeks from a stable (so i hear) :) 15:07
masak Nom-: perlgeek.de/blog-en/perl-5-to-6/ 15:08
Nom-: there have been stable releases of Parrot and Rakudo for a year and a half now. 15:09
15:09 rhr_ left
masak s/stable releases/stable monthly releases/ 15:11
however, the next release of Parrot will have a reasonably stable API. 15:12
15:12 gravity joined
szabgab Nom-, szabgab.com/perl6.html 15:13
have to go now, bye 15:14
masak szabgab: see you around.
15:14 ZuLuuuuuu joined
masak Nom-: it might also be good to know that we have a number of tools here on the channel that help elucidating the current state of the implementations. 15:21
perl6: say "hello Perl 6!"
p6eval elf 25239, pugs, rakudo 13f677: OUTPUTĀ«hello Perl 6!ā¤Ā»
15:22 shinobi_cl joined
masak that's three of our implementations replying. 15:22
std: class Lolipop {}; my Lolipop %hash; # a hash where the values are lolipops
p6eval std 25239: OUTPUTĀ«00:02 33mā¤Ā»
masak Nom-: STD knows whether things are legal Perl 6 or not. 15:23
Nom-: using either of these tools is usually a good way to strike up a conversation on the channel, because then people see what it is you're trying to acheive.
Nom- Fair enough :) 15:26
What would you recommend to start playing with ?
masak Nom-: among the implementations? Rakudo, and possibly Pugs.
Nom-: www.parrot.org/download 15:27
I recommend you to use Subversion to download Parrot.
Nom- I've actually fetched the 0.9.0 binary from there... would that come with Raduko ?
masak Nom-: it does, but not the very latest one. 15:28
however, nothing revolutionary has happened since the 0.9.0 release, so I guess that'll do.
if you do want the latest Rakudo, it's on github. but it requires the latest Parrot, not just 0.9.0. 15:29
Nom- I see... should I be seeing output errors for compile problems too?
masak Nom-: I'm not sure I understand the question. yes, I think so.
Nom- hmm... i've got a simple example from that blod you linked earlier, and it's not throwing any output... 15:31
I am pretty sure it's a compile problem, but something saying that would be nice :)
masak Nom-: sounds fishy.
Nom-: sial.org/pbot/perl6
you could paste the output here so that I can see what it is you're doing. 15:32
like this: sial.org/pbot/34936 15:33
Nom- sial.org/pbot/34937
masak gracias.
works for me. 15:34
Nom- hmm
masak might this be a Windows-related bug?
Nom- Entirely possible...
I'll build it for a real OS and see how I go :P
masak hey, anyone else on this channel run Windows? :) 15:35
care to confirm this bug for us?
I get this: sial.org/pbot/34938
Nom- right, which is what i'd expect 15:36
masak Nom-: at this point I would usually ask you to send the contents of your paste to [email@hidden.address] along with Parrot version information and a note that you're running Windows... 15:37
but this bug sounds very unlikely. someone else should have caught this already. 15:38
Nom-: there's a possibility that this is fixed in Parrot/Rakudo HEAD.
Nom- Certainly could be
I'm about to fire up a Ubuntu VM I should have lying around here... i'll try that out 15:39
masak Nom-: goodie.
but IMHO, it's bad enough if the latest Parrot release contains a Rakudo that doesn't print anything on Windows.
Nom- I only installed it on Windows because it was convenient :)
masak Nom-: thank you for your feedback so far.
it helps.
Nom- hmm i don't think I can get away with installing it on this VM... only 80MB memory free :P 15:40
masak that sounds a bit tight, yes. 15:41
Nom- It doesn't do much
Keeps my IRC idle time going, and receives email
That's about it
15:49 guru joined 15:57 finanalyst joined
Nom- 0.9.0 linux works perfectly for me 15:59
So it does appear to be Windows specific
masak Nom-: ok. if you want to, you can submit that bug report. 16:00
I don't know what else to do at this point. 16:01
I don't have Windows myself.
Nom- No probs, will do
masak Nom-++
ascent_ syntax error! =P 16:02
masak :) right, should be (Nom-)++
16:07 rhr_ joined
Nom- Hmm... does a database toolkit exist in parrot ? 16:08
masak Nom-: there have been experiments with sqlite bindings, but last I heard they were in their early stages. 16:10
Nom-: a few of us are writing a wiki engine in Rakudo Perl 6, and we're using plain files as out db backend right now.
Nom- I see... 16:11
I see a couple of blog posts about DBI 2.0 being based around the JDBC standard (when it eventually gets kicked off) 16:12
I think that's a bloody good idea :)
masak Nom-: it's a bit quiet here today. usually, we have more activity on #perl6. 16:13
Nom-: if you want to try an application written on top of Rakudo, do check out Druid.
github.com/masak/druid/
ascent_ what about CPAN? Something new planned? 16:16
masak ascent_: yes, but that was a long time ago. you can find some documents about it in the Pugs repository. 16:17
Nom- I was wondering that... where's CPAN6 :P
masak you're not the only one asking that question.
it's like with the mice. they all would like that bell to be hung around the cat, but... 16:18
...so far, nobody has actually gone ahead and done it.
16:19 rhr left 16:20 icwiener_ joined
ascent_ masak: so what are the requirements? I'll probably can do it. 16:23
masak: there was any longer discussion about it?
masak ascent_: kudos to you if you do it. it's a large project.
ascent_: I recommend reading up on that old docuement in the Pugs repo first.
16:25 nihiliad joined
masak I think this is it: svn.pugscode.org/pugs/misc/sixpan/ 16:28
someone correct me if I'n wrong.
16:29 ZuLuuuuuu left
ascent_ masak: thanks, greping pugs for cpan is a lot of... ;) 16:31
masak: I will look into it and check if I can handle it. 16:32
btw, what about threads in perl6?
masak ascent_: what about them? 16:35
16:35 Tene joined 16:36 icwiener left
ascent_ masak: I found that: "they've cut STM, and I suspect other threads related features may also be on the chopping block in the interest of delivering a GA language implementation in the next 18 months" 16:38
masak indeed.
that sounds like what jnthn said recently.
I don't recall why STM was cut; I kinda liked it. 16:39
ascent_ who cut that? I think threads are very important thing and it will be more and more important in future... my quadcore like 'em :) 16:42
masak ascent_: you know, junction autothreading is still in the language. 16:43
Nom- eep... so they actually did do STM and pulled it out? :(
masak Nom-: no, I don't think so.
Nom- awww, async NYI :( 16:46
16:47 Tene_ left
Nom- Is a Java style threading system on the books for perl6 ? 16:52
I'm very much looking forward to it, if it is
masak Nom-: svn.pugscode.org/pugs/docs/Perl6/Sp...rrency.pod 16:53
that's the latest, I think.
but even that should be outdated if STM was scrapped. 16:54
16:55 chrisdolan joined 17:02 shinobi_cl left
Nom- Well looks like none of that is implemented in rakudo 17:03
masak right.
however, a lot of other good stuff is. 17:04
perlgeek.de/blog-en/perl-6/rakudo-rocks.html :)
Nom- ooo.... perl 5.8.9 end of life :/ 17:06
We're only just about to make it off 5.6 @ work!
ascent_ masak: looks like someone is doing cpan already: cpan6.org ;)
masak well, I'll be. 17:07
nice page layout. too bad about the big table, though. 17:08
masak wonders how difficult it would be to do that the CSS way
Nom- Not difficult 17:09
CSS is never as hard as people make it out to be :P
masak likes CSS 17:10
one of the few standards in the world that actually hurts less the more you lean on it. :)
Nom- I do too.... CSS and XSLT are my two best friends
If XML wasn't so painful in Perl, i'd use XSLT more :) 17:11
masak I'm not sure what you mean by XML being painful in Perl. I don't think I know of any language where it's less painful. 17:12
Nom- It's painful NOT to use it in Java and .NET
Depends how you do it, but using Object Serialization is the easiest way to do XML 17:13
masak XPath is another wonderful standard, by the way. 17:15
Nom- Absolutely
masak I'm not sure I like XSLT so much. it's a bit too verbose and ambitious.
a lot of the shortcomings were fixed in XSLT 2.0, with the result that the spec is no longer readable by mortals. :/ 17:16
Nom- XSLT used in conjunction with object serialization in Java is the easiest way i've ever found to put a 'web interface' in front of an application
I use it all the time... have a java app spit some object out in XML format, and use XSLT to turn it into a human readable page
masak Nom-: I was a heavy Apache Cocoon user a few years back. XSLT does have its place, no doubt.
Nom- hmm i wonder if someone has made a perl 5.10 backport for Ubuntu Hardy ytet 17:17
17:18 kate21de joined 17:20 ayrnieu_ is now known as ayrnieu 17:22 rindolf joined
Nom- we've got like 1,000,000 lines of perl code which we're about to validate against Ubuntu Hardy, so if I could get Perl 5.10 for that, it's one less validation run when we go to the next LTS 17:23
masak sounds fun.
Nom- Perl 5.6.1 => Perl 5.10.0 ... I'd love to make that happen :) 17:25
dalek kudo: 6aaec33 | (Patrick R. Michaud)++ | (2 files):
Temporarily fix makefile construction to still work with Parrot.
pmichaud anyone around who can do commitbit magic? 17:33
17:34 guru left
diakopter pmichaud: me 17:46
for pugscode?
pmichaud: wait, what do you need
pmichaud: I SIGHUPed jifty on feather; commitbit.pugscode.org responds now... 17:48
pmichaud diakopter: thanks. 17:49
17:56 kate21de left
pugs_svn r25240 | chrisdolan++ | Add CDOLAN to author list 17:58
18:03 justatheory joined 18:07 kate21de joined
moritz_ re 18:10
18:11 justatheory left
dalek kudo: fb66ff6 | (Moritz Lenz)++ | t/harness:
[t/harness] fix testing if not checked out as languages/rakudo
 18:11
masak use.perl.org/~masak/journal/38430 18:18
nihiliad moritz_: re your t/harness commit: I'm sitting next to pmichaud at the Frozen Perl hackathon, and he asked me to fix the test harness so the tests will run as Perl 6 and not Perl 5. 18:20
moritz_: Are you already working on that?
moritz_ nihiliad: "make spectest" now works as it should - does that answer your question? 18:21
pmichaud I'm thinking we should completely rewrite t/harness 18:22
moritz_ that wouldn't be bad at all
18:22 moritz_ sets mode: +o pmichaud
moritz_ it's historically grown, as the saying goes ;-) 18:22
but whatever you do to the harness, please preserve its ability of parallel testing - that's increadibly useful to me 18:23
pmichaud well, currently t/harness tries to use Parrot's Parrot::Test::Harness, which in turn does a bunch of other things important to parrot testing.
I'm thinking we don't need Parrot::Test::Harness at all -- we should be just driving Test::Harness (from Perl 5) directly. 18:24
moritz_ aye
18:24 mncharity joined
pmichaud I'm also thinking that we need to put a 'parrot' symlink somewhere in the rakudo tree, so we aren't having to "guess" the version of Parrot that was used to build rakudo. 18:24
moritz_ nihiliad: sorry for stealing your task :/
s1n hey, pmichaud is back! 18:25
pmichaud well, the task of "fix current t/harness" is really different from "rewrite t/harness" :-)
moritz_ pmichaud: remember that some platfroms don't have good symlinks
(windows comes to mind)
mncharity pugscode.org/ planetsix.perl.org/ | nopaste: sial.org/pbot/perl6 | evalbot: perl6: say 3; (or rakudo:, pugs:, elf:, etc) | irclog: irc.pugscode.org/ | ~315 days 'til Xmas 18:26
moritz_ pmichaud: I thought about having a file in the rakudo repo that told us which parrot revision to use... 18:26
pmichaud: and a 'make parrot' target that updates parrot/ to that revision, and runs a 'make realclean; perl Configure.pl; make' there
18:27 maerzhase1 left
nihiliad moritz_: No worries. I'm a total n00b and will do whatever pmichaud tells me to. 18:27
pmichaud moritz_: there's already a file in the rakudo repo to do this (build/PARROTREV)
jnthn Windows sure doesn't do (good|any) symlinks...
s1n moritz_: why not just use parrot -V?
pmichaud I've thought about having a make target, but we have a chicken-and-egg problem there
i.e., we can't really build the Makefile until we know the parrot details 18:28
moritz_ pmichaud: we can have a bootstrapping makefile that gets overridden
pmichaud either that or we have to make the Makefile smart enough to work on .... yeah, something like that
moritz_ pmichaud: or offer that only as a convenience after the first run of Configure.pl
jnthn I guess I'm missing something but - why can't the Configure script work out where Parrot is and stick that location into the Makefile? 18:29
pmichaud moritz_: yes, I've been thinking about that also.
jnthn: that's what it does now.
jnthn: but we also need to know the location of parrot for other things, such as t/harness
jnthn OK. Is that a problem?
So make test just invokes t/harness with --parrot /path/to/parrot ? 18:30
s1n is it too much to assume it's on the user's $PATH?
moritz_ s1n: yes
s1n too soon? i mean, it's not like they can't install it
even from the repo
pmichaud "make install" doesn't work for Parrot right now. 18:31
jnthn: we're not always doing t/harness
sorry
we're not always invoking perl6 from 'make'
mncharity @tell TimToady Sorry I lost sync on STD. It turned out too costly. Staying with it would have preempted all other elf progress. I'll stick with rWhatever for now, and sync after bootstrap. Perhaps piecemeal syncs before then, not sure.
lambdabot Consider it noted.
jnthn Ah, true.
18:31 jdv79 joined
pmichaud e.g., if we run t/harness directly, or run tools/test_summary.pl -- they need to know the location of parrot 18:32
jdv79 szabgab: around?
szabgab jdv79, sorta, for a short time
jnthn Yes, it's rather burdensome to make the user pass Parrot's path whenever running those.
masak (Frozen Perl)++ # apparently, there's Rakudo hacking happening there!
food &
pmichaud yes, a lot.
jnthn I know it violates DRY, but Configure.pl could both put the path in the makefile and into a file, which t/harness etc can grab it from? 18:33
18:33 masak left
moritz_ well, for now Configure.pl could just write a file which contains the path to parrot 18:33
ah, jnthn had the same idea, and typed faster than me ;-)
jdv79 you realize Perl6::Conf is not straightforward to install, right?
szabgab jdv79, what do you mean ?
jnthn read outline of alester's talk 18:34
Sounds like we're in for a busy 2009! ;-)
szabgab for now I assumed PARROT_DIR and that rakudo is inside parrot
next version will require PARROT_DIR and RAKUDO_DIR
jdv79 yeah, that's a big piece
dalek kudo: 3d92f3f | (Francois Perrad)++ | (3 files):
improve .gitignore files
 18:35
kudo: 37e1492 | (Francois Perrad)++ | .gitignore:
ignore *.patch
pmichaud I thought about having a file with path to parrot, but I'd really like to come up with something cleaner
jdv79 can't we just use cpan versions of those so the deps can be done as usual?
pmichaud i.e., so that parrot can be run directly
szabgab jdv79, so what is the problem with Perl6::Conf installation ?
pmichaud instead of having all of the scripts read a file that says "where is parrot?" 18:36
s1n having a file with the path to parrot is no more easier or difficult than requiring parrot in their path
jdv79 its just a lot of work. it'd be nice if it "just worked"
jnthn pmichaud: To me, it's not perfect, but seems workable for now. Depends if you want a "works" kinda solution for a "beautiful" solution. :-)
mncharity moritz_/anyone: Could comment on a strawman concept? I'm thinking of creating a misc/elfish/gimme_elf/ with an elfg5 perl file. Where elfg5 is a standalone elf (basically elfblue with inlined src/perl6/STD.pmc for parsing - so no external parser). "Just download the one file (don't even have to svn co), install its cpan dependencies, and... instant taste of elf",
s1n i have parrot and perl6 in my repo builds already in my path to do just taht
pmichaud I need a solution that is easy to explain to others.
jnthn Nothing a lot better comes to mind rither off...
jdv79 i didn't know that rakudo moved out of languages/perl6 and to its own repo
mncharity Do you think that would excite folks? 18:37
pmichaud I can tell you that the number one problem right now at the hackathan is that people are completely unable to build parrot+rakudo and run perl6 without a _lot_ of explanation.
jnthn pmichaud: Easy to make others like, or easy to explain? ;-)
Ah.
szabgab pmichaud, I think you should remove languages/perl6 if that is not the place of rakudo any more
moritz_ mncharity: I think it's a very good idea (if it can be achieved with reasonable effort)
jdv79 pmichaud: i'm not even there and i just experienced that:(
18:38 Psyche^ joined, barney left
pmichaud szabgab: that's my intent -- we need to have a README there that says the new location and how to get the correct information 18:38
szabgab pmichaud, maybe replace with a README file that explains where ...
so I wont' finish my sentence :-)
s1n i just have scripts do the grunt work for me
jnthn Ah, sorry, gotta go now - back later.
moritz_ pmichaud: for now you can point to my blog at perlgeek.de/blog-en/perl-6/where-ra...writeback, or copy its information
mncharity moritz_: thanks... anyone else?
pmichaud the problem I'm having now is everyone says "we need this" but I'm too busy answering "then let's do XYZ" and not enough time to "do XYZ"
jnthn Happy hackathoning!
s1n maybe platform specific scripts or have the configure do something more
jdv79 szabgab: plus the POD doesn't render. it looks like p6 pod.
moritz_ pmichaud: or should I do it? 18:39
pmichaud moritz_: yes, I saw the blog post -- excellent.
moritz_: please do it.
anything that anyone can do such that I don't have to do it will save us all a ton of time. :-|
szabgab jdv79, well, yes it is p6 pod,
pmichaud right now I'm thrashing on "answering questions" and not getting any useful work done.
(not that answering questions isn't useful... :-)
szabgab I don't understand what did you expect from the AFAIK first real Perl 6 module on CPAN ?
18:40 Patterner left, Psyche^ is now known as Patterner
s1n what does the Configure.pl script do now anyways? 18:40
szabgab pmichaud, that's the sign of success
jdv79 nothing much really. i was hoping to be able to show it off but its not ready.
s1n it sounds like the job of the configure script to try and locate parrot 18:41
jdv79 nice work though
s1n at the very least, it could look in ../../, then in $PATH, then fall back on something else
pmichaud s1n: looking in $PATH isn't very useful at the moment. 18:42
currently Configure.pl looks at ../.. and parrot/
but a lot of the scripts and modules all assume ../..
s1n it's better than nothing
pmichaud and so we have to get them to not do that.
szabgab jdv79, when you setup CPAN.pm do you need to answer a few questions ?
pugs_svn r25241 | hinrik++ | [util/perl6.vim] allow non-keyword characters in the argument to =config
szabgab has anyone here managed to install the Perl6::Conf module using CPAN.pm ?
s1n just because parrot doesn't install properly, doesn't mean we can't add it to our path 18:43
jdv79 szabgab: it just broke cause of the PARROT_DIR req not working out
s1n i do that (export PATH=~/sandbox/parrot/:$PATH) already in my bashrc, maybe we can start making some assumptions to common setups
at least until parrot installs correctly
szabgab what do you mean not working out? you have not set it up or you set it up but it still could not install ? 18:44
s1n otherwise, you're writing a file the user has to modify to signal where to find parrot
pmichaud but the Configure.pl step can create that file
jdv79 i didn't retry with CPAN.pm. after i setup PARROT_DIR i did it by hand.
pmichaud ooooh, maybe parrot.pl ? ;-)
s1n and how is it going to create it? will it guess at parrot's location?
jdv79 szabgab: broke meaning the install failed.
pmichaud but that sounds pretty awful also, I guess. 18:45
s1n all solutions will suck until parrot correctly installs
szabgab jdv79, sure, there are tons of Perl5 modules with external dependencies that fail to install unless you configure that external dependency
s1n we can either start guessing where it will be, i.e. for common paths
or we make the user tell us 18:46
pmichaud even when parrot correctly installs, we still have to have a way to indicate the location of parrot
s1n no you don't
how do i know where gcc is?
my path
jdv79 szabgab: no doubt. but there is a parrot cpan dist. my question is can that be used.
pmichaud that's different
you only need gcc to build, not to execute
in rakudo's case, we need parrot both to build _and_ execute
s1n what about bash?
pmichaud what about bash? 18:47
s1n i think it's implied that it _has_ to be on your path
i need bash to execute things, it's in my path, so i don't have to say /bin/bash (crappy example, i know)
pmichaud it's implied that parrot has to be in one's path? I'm not sure that's a reasonable assumption at the moment.
s1n not right now it's not
but eventually
mncharity :/
s1n until then we can only guess or make the user tell us 18:48
%perl Configure.pl --parrot /path/to/parrot
pmichaud s1n: I already have that.
s1n or something else equally stupid
pmichaud you're solving the wrong problem.
s1n great, that's about as good as it'll get for now
what's the problem? lol
szabgab jdv79, AFAIK not yet as it does not really install
pmichaud or, more precisely, that's not currently sufficient
s1n care to elaborate? 18:49
pmichaud knowing the path to parrot is fine at build time -- I can do that.
szabgab as I can see the other chats they are just discussing this
pmichaud but after perl6.pbc is built, how do we invoke it?
jdv79 szabgab: i have a /usr/local/bin/parrot. it seems to not fetch and build rakudo though.
s1n why not use it to just build the perl6 executable?
pmichaud you mean the fakecutable?
because the perl6 executable often segfaults improperly. 18:50
s1n otherwise, you'll need a goofy path hack, file that specifies the path that is loaded, or something else
rakudo_svn r36451 | moritz++ | [rakudo] Rakudo left the nest - so remove most of languages/perl6/ and change
r36451 | moritz++ | the README accordingly.
szabgab jdv79, once there is a perl6 or rakudo executable in your PATH the installation can be straight forward
s1n okay, then Parrot.pm defines it?
then all the scripts use Parrot.pm, which defines the configured parrot info
szabgab till then we have to figure out the other things, eg. to had the pods look good
pmichaud I agree, the alternative is a goofy path hack. I'm not liking any of the solutions currently at our disposal, no.
s1n and we could generate a script that looks like perl6 that launches parrot from Parrot.pm 18:51
jdv79 szabgab: you mean getting search.cpan.org to use a 6 POD lib?
pmichaud "generate a script" -- how does that work under Windows?
jdv79 s/6/p6/
s1n that would have to be platform specific
szabgab jdv79, I am sending and e-mail to Graham now
pmichaud even platform specific -- how does that work under Windows?
you mean a perl6.bat ?
s1n or better yet, perl6.p6 :) 18:52
szabgab is there a perl6 pod parser written in perl 5 ?
moritz_ szabgab: yes, Damian wrote one
s1n perl6.p6 loads Parrot.pm, which defines parrot's location, executes it with the pbc
pmichaud szabgab: Damian had one at one time, but I don't know that it was ever released.
moritz_ it's on CPAN
pmichaud maybe we just should create a perl6 "executable" in C that does the work.
i.e., not the fakecutable approach -- something that just turns around and calls exec() 18:53
s1n naw, i like a perl6 script that executes the pbc :)
could do that, but that sounds like a p.i.t.a.
pmichaud all of the answers seem like a p.i.t.a at the moment.
szabgab Perl6::Perldoc::Parser thanks
s1n i actually think i like the idea of the perl6.p6 script that does the exec :)
pmichaud I don't.
s1n heh 18:54
pmichaud for one, that script can't be .p6
s1n well true
at least perl5
pmichaud so then running perl6 would mean that we start by invoking p5?
that seems like it'll get even more confusing.
s1n is there anything else that equates to something less than a pita? 18:55
pmichaud if there were an obvious simple solution it probably would've already been done. 18:56
s1n Parrot.pm is generated with parrot environment settings, perl6.pl (which is in the repo) loads Parrot.pm and does it's magic. all the scripts can then safely use Parrot.pm directly or through perl6.pl...
true, so we just need to find one that works until we can safely assume parrot is available already, right?
pmichaud "one that works" -- depends on the definition of "work" 18:57
s1n imho works means it requires the least work for a developer/user to get started
pmichaud it's not that we're choosing between one solution that works and many that don't -- it's that each solution has some downsides that aren't obvious
s1n if it's a pita for us, so be it
pmichaud I don't think perl6.pl is the answer I'm looking for. 18:58
I think I'd rather go with the fakecutable approach. It's more along the lines of what people will expect.
s1n how will you create that fakexcutable?
will it need to be compiled somehow?
pmichaud we currently do it with "make perl6" 18:59
s1n please tell me that won't introduce automake
oh
so just require the perl6 target to be executed?
i mean, that does a pbc_to_exe, which is problematic
pmichaud why is that problematic? 19:00
rakudo_svn r36453 | moritz++ | [rakudo] remove a left over perl6.pir file
s1n i thought you said it crashes intermitently and is kinda buggy 19:01
moritz_ it worked quite stable for me most of the time
pmichaud yes, but I see the perl6.pl as introducing a different set of issues. 19:02
19:02 legis left
pmichaud and yes, it's stable "most of the time" 19:02
maybe if it's a little intermittent we'll be pushed to get things fixed :-)
19:02 szabgab left
pmichaud but it's certainly far easier to say "perl6 ... " than anything else. 19:02
s1n how can the egg hatch itself without knowing about the chicken... hmmm
pmichaud: yes, it is, and that's why it's in my path :) 19:03
pmichaud I vote for the fakecutable approach.
s1n so don't change anything?
pmichaud well, we still need to fixup the harness
and we should change the default Makefile target to build the fakecutable 19:04
s1n i use the pbc right now, the perl6 segfault spews all kinds of crap
moritz_ currently 'make spectest' works for me.
pmichaud as long as 'make spectest' works with the fakecutable i'm reasonably okay with that approach.
I've also decided that I no longer want Parrot to be building Rakudo's Makefile 19:05
s1n good decision
pmichaud i.e., the "reconfigure.pl" approach.
(i.e., don't use Parrot's 'reconfigure.pl')
s1n pmichaud: are you at a hackathon right now? 19:06
pmichaud yes.
cognominal maddingue++
19:17 Tene_ joined
pugs_svn r25242 | putter++ | [gimme_elf] Created. A *very* easy to install, standalone, elfblue variant. To create interest in elf. 19:24
r25242 | putter++ | Sketched in README and Makefile.
mncharity moritz_: Thanks for the encouragement. :) 19:25
hmm, that commit should have been moritz_++. Wouldn't have happened without it.
g'day all &
19:25 mncharity left
moritz_ @tell mncharity now you have to blog about gimme_elf, otherwise the interest won't leave #perl6 19:26
lambdabot Consider it noted.
19:31 Tene left 19:35 FurnaceBoy joined 19:39 rindolf left 19:44 jferrero joined, rindolf joined
pugs_svn r25243 | hinrik++ | [util/perl6.vim] ensure proper nesting of angles inside Pod formatting codes 19:45
r25244 | hinrik++ | [S26] clarify the entry on E<> in the summary
r25245 | hinrik++ | [util/perl6.vim] update timestamp 19:46
r25246 | moritz++ | [irclog] point r\d+ on #parrot to parrot's trac installation 19:50
dalek kudo: 46f5a49 | (Moritz Lenz)++ | Configure.pl:
Apply coding style fix by Franōæ½xE7ois Perrad, fperrad++
 19:52
20:00 jhuni joined, rindolf left 20:03 maerzhase joined
chrisdolan I want to write a test for RT#62898, but I'm not sure where to put it. Advice? 20:05
dalek kudo: cb3422f | (Moritz Lenz)++ | .gitignore:
.gitignore `perl6'
 20:07
chrisdolan I don't think it's specced behavior, it's just implicit. Does it go in t/spectests or just t/?
moritz_ chrisdolan: into t/spec/ for sure
it's implicitly specced by basic sanity, and by "everything that's not specced works like in p5" 20:08
chrisdolan makes sense
moritz_ chrisdolan: maybe add it to t/spec/S12-class/basic.t
chrisdolan Cool, thanks
20:08 rindolf joined
moritz_ chrisdolan: do you know how I can easily pull your patch for that issue into the official rakudo tree? 20:10
chrisdolan I'm just learning git today... Let me ask here at the hackathon
20:11 riffraff joined
s1n so is pmichaud _the_ admin for the git repo now, or has he issued rights to someone else? 20:11
moritz_ chrisdolan: the patch you submitted contains spurious linebreaks
chrisdolan Yeah, that was my email client, sorry. I should have done it as an attachment. Let me research the git solution 20:12
moritz_ chrisdolan: one way is `git format-patch origin', but it's not the solutions I'm looking for ;-)
dalek kudo: 71179f1 | (Patrick R. Michaud)++ | (5 files):
Merge branch 'master' of [email@hidden.address]
 20:13
moritz_ s1n: I don't know - I don't think I can add new committers, but I can push changes for sure
20:13 M_o_C left
chrisdolan moritz_: OK, I just used the github.com "pull request" feature to send it to you 20:14
pmichaud so far I'm the admin for the git repo, yes.
moritz_ chrisdolan: thanks, got the message - but it doesn't contain instructions on how to pull :(
pmichaud I've generally granted push rights to those who were parrot committers
chrisdolan moritz_: Hrm... 20:15
moritz_ pmichaud: you seem to have missed fperrad (unless you added him in the last 30 minutes)
pmichaud probably so. I wasn't trying to be exhaustive
s1n what's the review process now to get push access? 20:16
pmichaud basically "if you're already a parrot committer, I'll add w/o question when needed"
s1n what about those who weren't parrot committers...
moritz_ chrisdolan: ah, `git-pull git://github.com/chrisdolan/rakudo.git master' seems to work
20:16 rindolf left
pmichaud I have to have some discussions with TPF and the like to see about committer access 20:16
chrisdolan moritz_: No, the patch isn't in my master branch. I made a branch called "package-redeclaration" for this 20:17
s1n that would be great, i'd like to know what it would take for me to get access
pmichaud we may still want to have the contributor license agreements in place
chrisdolan moritz_: so maybe it's just s/master/package-redeclaration/ from your previous command 20:18
moritz_ chrisdolan: I'll try that
chrisdolan: ah, that seems to really work, at least git-log shows your commit plus a merge commit 20:19
which is the git way of doing it, if I understand it correctly 20:20
chrisdolan sweet
20:21 jferrero left 20:22 rindolf joined
moritz_ rakudo: eval q[BEGIN {class Foo::Bar::Baz {}}; class Foo::Bar {}]; say "alive"; say $! if $!; 20:24
p6eval rakudo 6aaec3: OUTPUTĀ«No arguments found to dispatch onā¤current instr.: 'parrot;Perl6Role;!select' pc 2689 (src/classes/Role.pir:108)ā¤Ā»
moritz_ rakudo: eval q[BEGIN {class Foo::Bar::Baz {}}; class Foo::Bar {}]; say "alive";
p6eval rakudo 6aaec3: OUTPUTĀ«No arguments found to dispatch onā¤current instr.: 'parrot;Perl6Role;!select' pc 2689 (src/classes/Role.pir:108)ā¤Ā»
s1n is that revision the last 6 bytes of the index hash? 20:25
moritz_ chrisdolan: you could add that (with eval_lives_ok instead of eval) to S12-class/basic.t
s1n: it's the first 6 bytes of git-log | head -n 1
s1n: I don't know if that makes any sense :/
s1n moritz_: yeah sorta, git hashes still don't indicate timeline or anything, so it's as much sense as a sha1 hash can make 20:28
pugs_svn r25247 | chrisdolan++ | Add test for the class portion of: 20:29
r25247 | chrisdolan++ | RT#62898: Rakudo can't handle A declared after A::B was declared in an included module
chrisdolan Could someone commit the patch from rt.perl.org/rt3/Public/Bug/Display....?id=60164? 20:30
I got approval from pmichaud
moritz_ chrisdolan: I'm running a 'make spectest' now. If all passes, I'll push.
chrisdolan moritz_: still adding more tests...
moritz_ chrisdolan: ok, pushed. Many thanks. 20:33
dalek kudo: 2c026f7 | (Moritz Lenz)++ | src/parser/ (2 files):
Merge branch 'package-redeclaration' of git://github.com/chrisdolan/rakudo
 20:34
kudo: 1861489 | (Moritz Lenz)++ | .gitignore:
Merge branch 'master' of [email@hidden.address]
moritz_ chrisdolan++ # since it's nowhere in the commit message 20:35
pugs_svn r25248 | chrisdolan++ | A more specific test for RT#62898 20:37
20:42 jferrero joined, Limbic_Region joined 20:43 meppl left 20:57 IllvilJa left 21:12 jhuni left 21:18 finanalyst left 21:22 kate21de left
dalek kudo: 577566e | (Patrick R. Michaud)++ | (2 files):
More fixes to Configure.pl and makefiles
 21:23
kudo: 7ab01c5 | (Patrick R. Michaud)++ | (2 files):
Change Configure.pl to use build/Makefile.in.
moritz_ I'm starting to like the way git works 21:27
21:29 Chillance joined 21:30 riffraff left 21:31 kate21de joined 21:37 kate21de left 21:38 kate21de joined 21:40 rindolf left, Limbic_Region left
chrisdolan The rakudo portion of the frozen perl hackathon is breaking up now. 21:41
Thanks, moritz 21:42
21:50 kate21de left, chrisdolan left, kate21de joined 21:53 justatheory joined, rhr joined 21:55 FurnaceBoy left 22:05 icwiener-_- joined 22:06 ejs joined 22:07 rhr_ left 22:13 IllvilJa joined, nihiliad left 22:19 icwiener-_- is now known as icwiener 22:21 icwiener_ left 22:32 meppl joined 22:46 ejs left 22:54 DemoPhreak joined 23:12 DemoFreak left 23:24 elmex left, elmex joined 23:26 kate21de left, kate21de joined 23:28 DemoPhreak is now known as DemoFreak 23:30 DemoFreak left, iblechbot left 23:38 hercynium joined 23:40 maerzhase left
pugs_svn r25249 | particle++ | [S19] various minor updates and removal of stale comments 23:58

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!